CN101741786B - Ultra-broadband receiver for digital communication system and signal processing method thereof - Google Patents

Abstract

The invention discloses an ultra-broadband receiver for a digital communication system and a signal processing method thereof. The receiver comprises an ultra-broadband receiving module, a high-dimensional noise space signal demodulating module and a decoding judging module which are connected sequentially, wherein the high-dimensional noise space signal demodulating module comprises a plurality of parallel combing user signal collectors and soft information calculators correspondingly connected with the user signal collectors; outputs of all soft information calculators are connected to the decoding judging module; and the receiver firstly receives and samples a sending end signal and a disturbance noise signal through a receiving filter of the ultra-broadband receiving module, sends the acquired signals to the high-dimensional noise space signal demodulating module for further processing, outputs the soft judgment information to the decoding judging module and restores original data information. The receiver has the advantages of simple structure, convenience for hardware implementation and capability of remarkably improving the signal-to-noise ratio of the communication system. By adopting the receiver and the processing method under the same signal-to-noise ratio environment, the system can obtain performance improvement of not less than 0.25dB of error rate.

Description

Ultra-broadband receiver for digital communication system and signal processing method thereof

Technical field

The invention belongs to digital communication technology field; The signal processing that relates to receiving terminal in the digital communication system; For a kind of new digital receiver structural design and the processing method of disturbing noise cancellation signal, be specially a kind of ultra-broadband receiver for digital communication system and signal processing method thereof.

Background technology

In the various types of communication system, for radio communication,, signal of communication cause receiving terminal signal fadeout to occur especially owing to can receiving the channel transfer characteristic influence usually in transmission course; The opening of channel in addition, it makes that also aliasing has a large amount of ambient noises or other unknown signalings in the signal of communication of actual reception, these all cause harmful effect to the operate as normal of system.For this reason; In traditional receiver design and realizing; Usually before signal gets into receiver; The narrow band filter (its pass band width W is identical or bigger than communication bandwidth B, W ≈ B) that at first passes through front end with out-of-band noise and other unknown signaling filterings, guarantees that signal of communication can normally pass through with as much as possible simultaneously.According to this structure, people's design has obtained the general method for designing of optimum receiver, and draws the concrete form of receiver to the difference design of unlike signal and communication condition.

Fig. 1 has provided traditional receiver structure and handling process.Among the figure, signal of communication gets into before the receiver, and it at first will carry out Filtering Processing through a preposition narrow band filter, filtering out-of-band noise or other unwanted signals; Next, receiver is done further processing to filtering signal, comprises down-conversion, separates and be in harmonious proportion operation such as decoding judgement, recovers the primary data information (pdi) that is included in the reception signal at last.

After this, people have proposed multiple technologies such as multiplex technique, multi-antenna technology and iterative decoding again in succession, like code division multiple access (CDMA) system, orthogonal multiple carrier (OFDM) system, multiaerial system (MIMO) and smart antenna etc.These technological employings make the transmission characteristic of system obtain remarkable improvement; Yet make a general survey of present various types of communication system; Although it exists many difference at aspects such as system configuration, applied environment, access capability and sport technique segments; But be identical in essence; Promptly all the unwanted signal that comprises noise etc. that is mingled with in the communication channel is regarded as the harmful effect to communication system, and before it gets into receiver, utilizes foremost filter as much as possible with the unwanted signal filtering outside the passband.

Obviously, these for the intrinsic understanding of out-of-band noise or other unknown signalings but in limit more the design of receiver, also limited the further improvement of systematic function.

Summary of the invention

The problem that the present invention will solve is: in the existing communication system, outer noise or other unknown signalings of band all is regarded as harmful effect, and before it gets into receiving terminal, adopts narrow band filter to carry out filtering.These for the intrinsic understanding of out-of-band noise or other unknown signalings but in limit more the design of receiver; Also limited the further improvement of systematic function; Novel receiver structure need be provided and disturb the noise cancellation signal processing method; Make communication system can obtain higher-quality transfer of data, more effectively overcome the influence of noise and other unknown signalings communication system.

Technical scheme of the present invention is: ultra-broadband receiver for digital communication system; Comprise the ultra broadband receiver module, higher-dimension spatial noise signal demodulation process module and the decoding judging module that connect successively; The ultra broadband receiver module comprises receiving filter and high speed over-sampling device; The bandwidth of said receiving filter is far above the bandwidth of communication system transmitting terminal signal, and receiving filter exports high speed over-sampling device to, and the sampling rate of high speed over-sampling device is identical with the bandwidth of receiving filter; Higher-dimension spatial noise signal demodulation process module comprises a plurality of parallel branch comb formula subscriber signal gatherers; Each divides the output of comb formula subscriber signal gatherer to be connected with soft information calculator; The output of all soft information calculator is connected to the decoding judging module; Wherein dividing comb formula subscriber signal gatherer is the filter that is complementary with the transmitting terminal signal, and high speed over-sampling device exports said parallel branch comb formula subscriber signal gatherer to.

The receiving filter bandwidth of ultra broadband receiver module is more than 2 times of actual transmission signal bandwidth.

It is comb filter that the present invention preferably divides comb formula subscriber signal gatherer.

The signal processing method of above-mentioned ultra-broadband receiver for digital communication system; Receiving terminal in communication system; At first the receiving filter by the ultra broadband receiver module receives transmitting terminal signal and the outer noise cancellation signal of disturbing of transmitting terminal signal band in the lump, again the signal that receives is carried out over-sampling, and resulting oversampled signals is delivered to higher-dimension spatial noise demodulation module and done further processing; After receiving oversampled signals stream; The a plurality of parallel branch comb formula subscriber signal gatherer of higher-dimension spatial noise demodulation module is accomplished the multidiameter delay of signal and is handled, and declares information to deciphering judging module completion decoding judgement by soft information calculator output is soft, recovers original data message;

Over-sampling in the ultra broadband receiver module is:

It is that the transmission signal of B carries out broadband reception to bandwidth that the ultra broadband receiver module adopts receiving filter, and wherein B is for sending the bilateral spectral width of signal equivalence baseband signal, and the pass band width of receiving filter is W ^Sup, be the actual transmission signal bandwidth B K doubly, i.e. W ^Sup=KB, K>=2; Next, high speed over-sampling device carries out over-sampling to filter output signal, and sampling rate does

The corresponding sampling interval

It is standard nyquist sampling interval T _SThe 1/K of=1/B, promptly

If the mark space of signal of communication is T ₀, then signal sampling point can be written as on l mark space:

$r_l\left(p{T}_S^{\mathrm{Sup}}\right)={\mathrm{\α}}_l\·s_l\sqrt{E_S}g(p\·T_S^{\mathrm{Sup}})+n^{\′}(p\·T_S^{\mathrm{Sup}})$

Wherein, p representes sample number, p=0 ..., K [T ₀/ T _S]-1 is promptly at mark space T ₀On adopt total sample be K [T ₀/ T _S], α _lBe the channel fading factor, s _l∈ Φ is a certain data symbol that l time slot sends, and Φ is data symbol s _lMight value all, E _SBeing the energy of single symbol, is W for pass band width ^SupFilter, the noise signal component n ' of its output (t) is uniformly-spaced putting t=p/W ^SupThe sampling point at place

Be the Gaussian random variable of separate same distribution, average is 0, and variance is N ₀/ 2, N ₀/ 2 is the bilateral power spectral density of noise;

According to sampling rate The oversampled signals sample that obtains after the sampling

Delivering to higher-dimension spatial noise signal demodulation module handles; Higher-dimension spatial noise signal demodulation module correspondence is provided with K parallel branch comb formula subscriber signal gatherer; Collect and screening operation in order to accomplish useful signal, divide comb formula subscriber signal gatherer according to different separately extraction of example side-play amounts respectively to oversampled signals

Carry out extracted at equal intervals, and carry out parallel processing to extracting signal, for the branch comb formula subscriber signal gatherer that is numbered k, branch number of living in is k, and corresponding sample extracts side-play amount and does $k\·T_S^{\mathrm{Sup}}(k=0,...,K-1),$ It extracts the result:

$r(q;k)\stackrel{\mathrm{\Δ}}{=}{r}_l(q\·T_S+k\·T_S^{\mathrm{Sup}})$

$={\mathrm{\α}}_l\·s_l\sqrt{E_S}g(q{T}_S+k{T}_S^{\mathrm{Sup}})+n(q{T}_S+k{T}_S^{\mathrm{Sup}})$

Wherein, q=0 ..., [T ₀/ T _S]-1 is for extracting the numbering of sample value, the extraction sampling point that following formula obtains is actual be original signal r (t) with

As its initial offset time, according to standard nyquist sampling interval T _SThe result who samples is known r (q by sampling thheorem; K) can recover primary signal r (t) equally, the useful signal component that each road divides the extraction result of comb formula subscriber signal gatherer to be comprised is of equal value fully, and has undistortedly:

$\underset{q}{\mathrm{\Σ}}{\left|g(q{T}_S+k_i\·T_S^{\mathrm{Sup}})\right|}^2=\underset{q}{\mathrm{\Σ}}{\left|g(q{T}_S+k_j\·T_S^{\mathrm{Sup}})\right|}^2=1$

Be that the useful signal energy that each minute comb formula subscriber signal gatherer is collected all equates, and equal to send signal energy, divide comb formula subscriber signal gatherer k to use and extract signal point range { r (q; K) } _qIn the pulse that is complementary of useful signal

Accomplish and extract the signal filtering reception, its output result is:

$d_l\left(k\right)=\⟨r_l(q\·T_S+k{T}_S^{\mathrm{Sup}}),g(q\·T_S+k{T}_S^{\mathrm{Sup}})\⟩$

$={\mathrm{\α}}_l\·s_l\sqrt{E_S}\⟨g(q;k),g(q;k)\⟩+\⟨n(q;k),g(q;k)\⟩$

$={\mathrm{\α}}_l\·s_l\sqrt{E_S}+n_l\left(k\right)$

Wherein, is for dividing the filtering output noise component of comb formula subscriber signal gatherer; Be the Gaussian random variable of zero-mean, this variable corresponding variance is:

${\mathrm{\σ}}_{n^{\′}}^2=\underset{q}{\mathrm{\Σ}}[E{\left|n(q;k)\right|}^2\·g(q;k)g^{*}(q;k)]$

$=\frac{N_0}{2}\·\left\{\underset{q}{\mathrm{\Σ}}{\left|g(q;k)\right|}^2\right\}=\frac{N_0}{2}$

So, the output noise component n of branch comb formula subscriber signal gatherer k _l(k) probability density function profiles is:

$p[n_l\left(k\right)=n]=\frac{1}{\mathrm{\π}{N}_0/2}\exp ({\left|n\right|}^2/(N_0/2))$

Wherein, n _l(k)=n representes the numerical value of the filtering output noise component of current branch comb formula subscriber signal gatherer, this moment d _l(k) conditional probability density also can correspondingly obtain:

$p\left(d_l\left(k\right)\right|s_l,{\mathrm{\α}}_l)=\frac{1}{\mathrm{\π}{N}_0/2}\exp (-{|d_l\left(k\right)-{\mathrm{\α}}_l\·s_l|}^2/(N_0/2))$

Next, the corresponding soft information calculator that connects of branch comb formula subscriber signal gatherer k can be according to the d of current output _l(k) and the channel fading factor-alpha _l, calculate the data symbol s of transmission _lGet a certain symbol of set Φ

m ₁=0 ..., the soft value of declaring during M-1:

$P_k(s_l=s_{m_1})=\frac{p\left[d_l\left(k\right)\right|s_{m_1},{\mathrm{\α}}_l]}{\underset{s_{m^{\′}}\∈\mathrm{\Φ}}{\mathrm{\Σ}}p\left[d_l\left(k\right)\right|s_{m^{\′}},{\mathrm{\α}}_l]}$

The system number that the communication system of representing M adopts, m ₁Be symbol number, s _{M '}Be the possible value of the institute of set Φ, following formula is illustrated in current channel condition α _lDivide comb formula subscriber signal gatherer k to receive signal d down, _l(k) judgement maybe value for symbol

Probability, existing with d _l(k) conditional probability density substitution following formula, can obtain branch road k output about data symbol s _lThe soft value of the declaring calculating formula of all feasible solutions:

$P_k(s_l=s_{m_1})=\frac{\exp ({|d_l\left(k\right)-{\mathrm{\α}}_l\·s_{m_1}|}^2/(N_0/2))}{\underset{s_{m^{\′}}\∈\mathrm{\Φ}}{\mathrm{\Σ}}\exp ({|d_l\left(k\right)-{\mathrm{\α}}_l\·s_{m^{\′}}|}^2/(N_0/2))},m_1=0,...,M-1$

Then, branch road k will obtain about data symbol s _lSoft value of declaring of each feasible solution

Deliver to decoding decision device module, and obtain the soft value of declaring together with other parallel branch and accomplish last symbol judgement, recover raw data symbols.

Further, the decoding judging module selects the maximum a posteriori criterion as decision rule, earlier will by higher-dimension spatial noise signal demodulation process module obtain about data symbol s _lSoft value of declaring of each feasible solution add up respectively, that is:

$P(s_l=s_{m_1})=\underset{k=0}{\overset{K-1}{\mathrm{\Σ}}}[P_k(s_l=s_{m_1})/\underset{m_1=0}{\overset{M-1}{\mathrm{\Σ}}}\underset{k=0}{\overset{K-1}{\mathrm{\Σ}}}{P}_k(s_l=s_{m_1})]$

Then selecting wherein, the pairing symbol value of maximum recovers raw data symbols as the optimal judgement result:

${\hat{s}}_l=\underset{s_{m_1}\∈\mathrm{\Φ}}{\arg }\left\{\underset{m_1=0,...,M-1}{\max }P(s_l=s_{m_1})\right\}.$

Among the present invention; At first adopt the receiving filter of ultra broadband receiver module to substitute the preposition narrow band filter of existing system at receiving terminal; The bandwidth of receiving filter is several times of actual transmission signal bandwidth, and corresponding raising receives signals sampling speed, and guarantees that the statistics between the sampled point is independent; Next, at higher-dimension spatial noise demodulation module, it adopts and parallel divides a comb formula subscriber signal gatherer that high-speed data-flow is transformed to carry out uniformly-spaced T according to different separately side-play amounts _S(=1/F _S) extract, thereby the rate data streams that converts multidiameter delay into is carried out parallel demodulation and processing.Therefore, each divides the noise component(s) n of comb formula subscriber signal gatherer output _l(k) independently of one another, therefore, the noise gross energy that K branch comb formula subscriber signal gatherer introduced is K (N ₀/ 2), be traditional narrow filter noise intensity K doubly; But on the other hand, the signal summation of being collected by each minute comb formula subscriber signal gatherer is also corresponding to have increased K doubly, and corresponding signal energy is according to K ²Multiplication is long.That is to say that though noise intensity of the present invention is K times of traditional approach, signal energy is the K of traditional approach ²Doubly, as a whole, the advantages outweigh the disadvantages, compares traditional approach, introduced extra reception processing gain, improved the signal to noise ratio of communication system.

The present invention can significantly improve the signal to noise ratio of communication system, and receiver structure of the present invention is simple, is convenient to hardware and realizes; Experimental result shows, under identical signal to noise ratio environment, adopts receiver of the present invention and processing method can make system obtain to be not less than the bit error rate performance improvement of 2.5dB.

Description of drawings

Fig. 1 is traditional receiver structure and handling process sketch map.

Fig. 2 is the regulation diagrammatic sketch of the present invention to the signal of communication bandwidth.

Fig. 3 is receiver structure of the present invention and handling process sketch map.

Fig. 4 is the sketch map of ultra broadband receiving filter of the present invention.

Fig. 5 is the sketch map of higher-dimension spatial noise signal demodulation process module of the present invention.

Fig. 6 is the performance test figure of receiver of the present invention under digital baseband system flat fading wireless channel.

Fig. 7 is the performance test figure of receiver of the present invention under ofdm system flat fading wireless channel.

Embodiment

Receiver is one of core component of Communication System Design; The objective of the invention is to propose a kind of novel ultra broadband noise introduction-type digital receiver structure; In order to transmission rate and the reliability of improving communication system, to support the transfer of data of higher rate.

In concrete the realization; The present invention is based on basis idea of Genetic Algorithm, adopt the ultra broadband receiving filter to substitute original preposition narrow band filter, undesirable out-of-band noise component is initiatively introduced receiver originally; And pass through high-speed sampling and on the superelevation dimension space, accomplish a series of special processings such as parallel signal collection and soft decoding judgement; Therefrom obtain processing gain, make the error rate descended significantly, effectively improved the transmission performance of system.

The present invention's unique distinction the most just is it and changed in the past the mode of " blocking up ", but initiatively will more be with outer unwanted signal component to introduce into receiver, and in order to participating in accomplishing resuming work of data message, and obtained performance gains in this processing procedure.Experimental result proves that this ultra broadband noise introduction-type receiver structure can effectively improve the transmission rate of system, and has favorable applicability, is applicable to existing various types of communication system.

Further specify in the face of the application of the present invention in digital communication system down.

1. digital communication system is described

If mark space is T ₀, the energy of single symbol is E _S, g (t) is for sending the pulse shaping function that symbol adopted, length T ₀, then originating terminal signal can be written as:

$s\left(t\right)=\sqrt{E_S}\·\underset{l}{\mathrm{\Σ}}{s}_{l}g(t-l\·T_0)---\left(1\right)$

Wherein, s _lBe the data symbol that l time slot sends, its value is counted M by modulation system and system and is determined, for binary system bipolarity amplitude-shift keying modulation 2ASK, then s _l∈ { ± 1}; If adopt multi-system amplitude-shift keying modulation M-PSK, then s _l∈ exp (j2 π m/M) | m=0 .., M-1}.More general, with data symbol s _lMight value all be called glossary of symbols Φ, i.e. s _l∈ Φ.Signal s (t) shared bandwidth B in process of transmitting is mainly determined by pulse shaping function g (t).It should be noted that the bandwidth of being mentioned in the literary composition refers to the bilateral spectral width of the corresponding equivalent baseband signal of signal of communication, specifically referring to Fig. 2 in order to discuss and uniformity of treatment.

In the actual transmissions process, because channel transfer characteristic is undesirable, the particularly influence of channel multi-path and time-varying characteristics, can there be decline in various degree usually in received signal, even comparatively serious intersymbol interference occurs.In addition, noise that is mingled with in the channel and unknown signaling also can cause the signal of communication distortion, influence the correct judgement of data symbol at receiving terminal.For the purpose of discussing conveniently; Below with additive white Gaussian noise channel AWGN and the time be example under the flattened fading channel condition; The operation principle of introducing system and realization, conclusion that this paper is given and method can directly expand to other communication system or more complicated channel circumstance.

Be n (t) if receiving terminal noise and unknown signaling disturb, be made as the zero-mean Gaussian random process, in time, becomes the fading channel Channel Transmission factor and is α (t), then receives signal and is:

$r\left(t\right)=\mathrm{\α}\left(t\right)\·s\left(t\right)+n\left(t\right)$

$=\mathrm{\α}\left(t\right)\·[\sqrt{E_S}\·\underset{l}{\mathrm{\Σ}}{s}_{l}g(t-l\·T_0)]+n\left(t\right)---\left(2\right)$

Here remember symbol W _NBe the bandwidth of noise n (t), N ₀/ 2 is the bilateral power spectral density of noise n (t).Generally, the bandwidth of interchannel noise and unknown signaling n (t) will be sent the bandwidth of signal, i.e. W much larger than system _N＞＞B.Be without loss of generality, especially under high speed data transfer, at mark space T ₀Interior channel time variation can be ignored, and only change at random possibly take place the upper signal channel state at the adjacent-symbol interval.So receiving signal can be abbreviated as:

$r\left(t\right)=\mathrm{\α}\left(t\right)\·s\left(t\right)+n\left(t\right)$

$=\sqrt{E_S}\underset{l}{\mathrm{\Σ}}{\mathrm{\α}}_l\·s_{l}g(t-l{T}_0)+n\left(t\right)---\left(3\right)$

Wherein, l representes l sending time slots, coefficient For sending symbol s _lThe channel fading that in transmission course, experiences is called fading factor, and it keeps constant in mark space; Correspondingly, we claim that the channel of this moment is a flat fading channel.For awgn channel, fading factor α (lT then ₀) be constant 1, α (lT is arranged ₀)=α=1,

2. traditional receiver structure and design principle

In traditional receiver, for fear of or reduce of the influence of extraneous unwanted signal as far as possible to receiver, at receiver front end a narrow band filter is set usually, its pass band width W _SSignal bandwidth B is identical or bigger with sending, W _S≈ B, its effect is to make useful signal normally pass through, simultaneously with unwanted signal filterings such as out-of-band noises.Next, receiver carries out bandpass sampling to filtered narrow band signal, and sampling rate is F _S=W _S, T _S=1/F _SBe the sampling interval of correspondence, then adopt on l mark space the signal sampling point be:

$r_l(c\·T_S)=r\left(t\right){|}_{t=(l-1)T_0+q\·T_S},(c=0,...,[T_0/T_S]-1)$

$={\mathrm{\α}}_l\·s_l\sqrt{E_S}g\left(c{T}_S\right)+n\left(c{T}_S\right)---\left(4\right)$

C is signal sampling point numbering, at this moment, and at this mark space T ₀On adopt total sample for [T ₀/ T _S], [.] is for rounding operator. For pulse shaping function g (t) at t=cT _S(c ∈ [0, [T ₀/ T _S]-1]) sample value, and have

Then as shown in Figure 1, sample sequence obtains being used to decipher the signal of judgement through processing such as down-conversion, coherent receptions:

$d_l={\⟨r_l\left(c{T}_S\right),g^{*}\left(c{T}_S\right)\⟩}_c={\mathrm{\α}}_l\·(\sqrt{E_S}\·s_l)+n_l---\left(5\right)$

Wherein,<.; Be inner product operator, n _lBe the noise component(s) of coherent reception output, be similarly the stochastic variable of Gaussian distributed, average is 0, variance

Equal:

${\mathrm{\&sigma;}}_n^2=E[T_S\underset{c_1}{\mathrm{\&Sigma;}}\underset{c_2}{\mathrm{\&Sigma;}}n\left(c_1{T}_S\right)n^{*}\left(c_2{T}_S\right)\&CenterDot;g\left(c_1{T}_S\right)\&CenterDot;g^{*}\left(c_2{T}_S\right)]$

$=N_0/2\&CenterDot;\left(\underset{c_1}{\mathrm{\&Sigma;}}{\left|g\left(c_1{T}_S\right)\right|}^2{T}_S\right)$

$=N_0/2---\left(6\right)$

More general, for the multiple Gaussian random variable n of circumference symmetry _l=real (n _l)+jimag (n _l), real part real (n _l) and imaginary part imag (n _l) all obey the Gaussian distribution of zero-mean and separate, its variance respectively equals

Half the

Then receive decision signal d _lThe conditional probability density of=d is:

$p(d_l=d|s_l,{\mathrm{\&alpha;}}_l)=\frac{1}{\mathrm{\&pi;}{N}_0/2}\exp (-{|d-{\mathrm{\&alpha;}}_l\&CenterDot;s_l|}^2/(N_0/2))---\left(7\right)$

Decision signal d according to output _lWith current channel condition α _l, the decoding decision device can be according to the make a start posterior probability (providing the result of calculation under the general condition such as symbol here) of each symbol of computes:

$P\left(s_l\right|d_l=d,{\mathrm{\&alpha;}}_l)=\frac{P(s_l,d|{\mathrm{\&alpha;}}_l)}{\underset{s_{l^{\&prime;}}\&Element;\mathrm{\&Phi;}}{\mathrm{\&Sigma;}}P(s_{l^{\&prime;}},d|{\mathrm{\&alpha;}}_l)}=\frac{P\left(d\right|s_l,{\mathrm{\&alpha;}}_l)}{\underset{s_{l^{\&prime;}}\&Element;\mathrm{\&Phi;}}{\mathrm{\&Sigma;}}P\left(d\right|s_{l^{\&prime;}},{\mathrm{\&alpha;}}_l)}---\left(8\right)$

According to symbol s _lMight value result of calculation in formula (8) size, the pairing symbol of decoding decision device selection posterior probability maximum As the court verdict of the best, that is:

${\hat{s}}_l=\underset{s_l}{\arg }\left\{\underset{s_l\&Element;\mathrm{\&Phi;}}{\max }\right[P\left(d\right|s_l,{\mathrm{\&alpha;}}_l)/\underset{s_{l^{\&prime;}}\&Element;\mathrm{\&Phi;}}{\mathrm{\&Sigma;}}P\left(d\right|s_{l^{\&prime;}},{\mathrm{\&alpha;}}_l)\left]\right|\}---\left(9\right)$

Wherein, We claim to decipher decision device and correctly adjudicate when court verdict , recover primary data information (pdi); And we claim to decipher decision device misjudgement have taken place at this moment when court verdict

, and the probability that the misjudgement incident occurs just is called the error rate or the error sign ratio of system.For communication system, the final purpose of its design reduces error probability exactly as much as possible under given speed and condition of work, improves the reliable transmission performance of system.

◆ for binary system bipolarity amplitude-shift keying modulation 2ASK, E _S=E _b, error sign ratio (theoretical value) under flattened rayleigh fading channel condition is as follows when slow at awgn channel for it:

$P_e^{2\mathrm{ASK}}{|}_{\mathrm{AWGN}}=\frac{1}{2}\mathrm{erfc}\left(\sqrt{{\mathrm{\&gamma;}}_b}\right),({\mathrm{\&gamma;}}_b=E_b/N_0)---\left(10\right)$

$P_e^{2\mathrm{ASK}}{|}_{\mathrm{Rayleigh}-\mathrm{fading}}=\frac{1}{2}(1-\sqrt{\frac{{\stackrel{\&OverBar;}{\mathrm{\&gamma;}}}_b}{1+{\stackrel{\&OverBar;}{\mathrm{\&gamma;}}}_b}}),({\stackrel{\&OverBar;}{\mathrm{\&gamma;}}}_b={\mathrm{\&gamma;}}_b\&CenterDot;E\left({\left|\mathrm{\&alpha;}\right|}^2\right))---\left(11\right)$

Wherein, E _bBe bit energy, ratio γ _b=E _b/ N ₀The bit signal to noise ratio of representing this system.E (| α | ²) be the power gain (statistical value) of fading channel,

For arrive the average bit signal to noise ratio of receiving terminal through fading channel transmission back.

◆ and, get M=4 here for multi-system phase keying modulation MPSK, comprise two bits in each symbol, at this moment E _S=2E _b, it is respectively at awgn channel error sign ratio (theoretical value) under the flattened rayleigh fading channel condition when slow:

$P_e^{4\mathrm{PSK}}{|}_{\mathrm{AWGN}}=\mathrm{erfc}\left(\sqrt{{\mathrm{\&gamma;}}_b}\right)[1-\frac{1}{4}\mathrm{erfc}\left(\sqrt{{\mathrm{\&gamma;}}_b}\right)]---\left(12\right)$

$P_e^{4\mathrm{PSK}}{|}_{\mathrm{Rayleigh}-\mathrm{fading}}=1-\frac{1}{4}{(1-\sqrt{{\stackrel{\&OverBar;}{\mathrm{\&gamma;}}}_b/(1+{\stackrel{\&OverBar;}{\mathrm{\&gamma;}}}_b)})}^2---\left(13\right)$

3. ultra wideband receiver structure of the present invention and realization

System compares with existing communication; Novel receiver proposed by the invention has adopted new design idea; Adopt the pass filter of ultra broadband to substitute the preposition narrow band filter among Fig. 1; The bandwidth of receiving filter is several times of actual transmission signal bandwidth, makes a large amount of out-of-band noises and unknown signaling component get into receiver, and obtains the structure and the signal processing method of novel receiver according to this design.It realizes that mainly by the ultra broadband receiver module three key components of higher-dimension spatial noise signal processing module and decoding judging module constitute, and are as shown in Figure 3, the bandwidth W of the receiving filter of ultra broadband receiver module ^SupFar above the bandwidth B of actual transmission signal, W ^Sup＞＞B generally gets more than 2 times.

Its handling process is following: at first adopt receiving filter that useful signal and the outer noise cancellation signal of disturbing of band are received in the lump, and signal is carried out the high speed over-sampling handle; Then, the high speed oversampled signals will be sent to higher-dimension spatial noise demodulation module and do further processing.Behind the sampled signal stream of receiving at a high speed, higher-dimension spatial noise demodulation module is accomplished the multidiameter delay computing and the processing of signal, and will export and softly declare information and deliver to decision device and accomplish final decoding and adjudicate, and recovers original data message.The design that provides the ultra wideband receiver each several part below realizes.

3.1. the structure of ultra broadband receiver module

The noise in the reality and the spectrum component of unknown signaling are very abundant, its frequency bandwidth W _NThan the signal of communication bandwidth B and Yan Yaokuan many, be designated as W here _N＞＞B.The effect of ultra broadband receiving filter will introduce receiver with these unwanted signals (comprising noise) that is positioned at beyond the communication band just.In the present invention, adopt pass band width W ^Sup=KB, K>=2 receiving filters and corresponding with it high speed over-sampling device are accomplished this task, see Fig. 4.

The sampling rate of note high speed over-sampling device does

Then

Wherein, K=[W ^Sup/ B]; The sampling rate that is to say among the present invention to be adopted

Be legacy communications system sampling rate F _SK doubly.And the span of factor K is by the frequency range W of interchannel noise _NRatio [W with signal bandwidth B _N/ B] determine, that is:

2≤K≤[W _N/B] （14）

Corresponding high-speed sampling at interval

Be reduced to original system sampling rate T this moment _S1/K, promptly

Therefore signal sampling point can be rewritten as on l the mark space drawing of formula (4):

$r_l\left(p{T}_S^{\mathrm{Sup}}\right)={\mathrm{\&alpha;}}_l\&CenterDot;s_l\sqrt{E_S}g(p\&CenterDot;T_S^{\mathrm{Sup}})+n^{\&prime;}(p\&CenterDot;T_S^{\mathrm{Sup}})---\left(15\right)$

Wherein, p=0 ..., K [T ₀/ T _S]-1, expression sample numbering is promptly in the same-sign interval T ₀On adopt total sample be K [T ₀/ T _S].For pass band width W ^SupThe receiving filter of=KB, the noise signal component n ' of its output (t) is uniformly-spaced putting t=p/W ^SupPlace's sampling point is separate, so the noise signal sample in the formula (15)

Be independent identically distributed Gaussian random variable, average is 0, and variance is N ₀/ 2.

At this moment, the receiving filter through the ultra broadband receiver module gets into the noise signal power of receiver and is:

$P_N=\frac{1}{T_0}\underset{p=0}{\overset{[T_0/T_S^{\mathrm{Sup}}]-1}{\mathrm{\&Sigma;}}}E\left[{\left|n^{\&prime;}(p\&CenterDot;T_S^{\mathrm{Sup}})\right|}^2\right]=\frac{n_0}{2T_0}\&CenterDot;[T_0/T_S^{\mathrm{Sup}}]$

$=\frac{N_0}{2}\left(F_S^{\mathrm{Sup}}\right)$

$=\frac{N_0}{2}\left(W^{\mathrm{Sup}}\right)---\left(16\right)$

Hence one can see that, and the noise average power that gets into receiver is directly proportional with filter bandwidht, and obviously its noise power than legacy system has been amplified K doubly, explains that more out of band signal gets into receiver.Next; The sample of signal that the high speed over-sampling obtains

will be sent to higher-dimension spatial noise signal processing module and handle, and export the soft information of declaring.

3.2. higher-dimension spatial noise signal demodulation process module

Higher-dimension spatial noise signal processing module is the another critical component of receiver of the present invention, is used for handling the outer unwanted signal of a large amount of bands that the high speed oversampled signals comprises, and its structure is formed as shown in Figure 5.Higher-dimension spatial noise signal demodulation process module comprises the branch comb formula subscriber signal gatherer of multidiameter delay; Each divides the output of comb formula subscriber signal gatherer to be connected with soft information calculator; Obtain through soft information calculator that the data symbol of each minute comb formula subscriber signal gatherer is soft declares information; These are soft declares information and will deliver to follow-up decoding judging module, for it provides necessary criterion.Wherein dividing comb formula subscriber signal gatherer is the filter that is complementary with the transmitting terminal signal; Preferred comb filter, it is little that comb filter is calculated the calculation amount, and speed is fast; Amount of calculation when helping reducing parallel processing, high speed over-sampling device export said parallel branch comb formula subscriber signal gatherer to.

As shown in Figure 5; The oversampled signals sample

that obtains behind the process high speed over-sampling is at first delivered to the branch comb formula subscriber signal gatherer of multidiameter delay, accomplishes useful signal and collects and screening operation.Here be provided with K parallel branch altogether, K=[W ^Sup/ B].Divide comb formula subscriber signal gatherer respectively oversampled signals

to be carried out extracted at equal intervals, and carry out parallel processing extracting signal according to different separately extraction of example side-play amounts.

$r_l(p\&CenterDot;T_S^{\mathrm{Sup}})={\mathrm{\&alpha;}}_l\&CenterDot;s_l\sqrt{E_S}g(p\&CenterDot;T_S^{\mathrm{Sup}})+n(p\&CenterDot;T_S^{\mathrm{Sup}})---\left(17\right)$

For the branch comb formula subscriber signal gatherer that is numbered k; Branch number of living in is k, and it is designated as r (q to the extraction result who crosses sample of signal to its extraction of example side-play amount for

; K):

$r(q;k)\stackrel{\mathrm{\&Delta;}}{=}{r}_l(q\&CenterDot;T_S+k\&CenterDot;T_S^{\mathrm{Sup}})$

$={\mathrm{\&alpha;}}_l\&CenterDot;s_l\sqrt{E_S}g(q{T}_S+k{T}_S^{\mathrm{Sup}})+n(q{T}_S+k{T}_S^{\mathrm{Sup}})---\left(18\right)$

Wherein, q=0 ..., [T ₀/ T _S]-1 is for extracting the numbering of sample value, the extraction sampling point that following formula obtains is actual be original signal r (t) with

As its initial offset time, according to standard Nyquist Nyquist sampling interval T _SThe result who samples is known r (q by sampling thheorem; K) can recover primary signal r (t) equally, the useful signal component that each road divides the extraction result of comb formula subscriber signal gatherer to be comprised is of equal value fully, and has undistortedly:

$\underset{q}{\mathrm{\&Sigma;}}{\left|g(q{T}_S+k_i\&CenterDot;T_S^{\mathrm{Sup}})\right|}^2=\underset{q}{\mathrm{\&Sigma;}}{\left|g(q{T}_S+k_j\&CenterDot;T_S^{\mathrm{Sup}})\right|}^2=1---\left(19\right)$

Be that the useful signal energy that each minute comb formula subscriber signal gatherer is collected all equates, and equal to send signal energy.And on the other hand; Know by formula (15); The different comb formula subscriber signal gatherer k=0 that divide; ..., it is independently of one another that K-1 extracts the noise component(s) that signal comprised, and this makes the improvement of performance become possibility.Next, divide comb formula subscriber signal gatherer k to use and extraction signal point range { r (q; K) } _qIn the pulse that is complementary of useful signal

Accomplish and extract the signal filtering reception, its output result is:

$d_l\left(k\right)=\&lang;r_l(q\&CenterDot;T_S+k{T}_S^{\mathrm{Sup}}),g(q\&CenterDot;T_S+k{T}_S^{\mathrm{Sup}}){\&rang;}_c$

$={\mathrm{\&alpha;}}_l\&CenterDot;s_l\sqrt{E_S}\&lang;g(q;k),g(q;k)\&rang;+\&lang;n(q;k),g(q;k)\&rang;$

$={\mathrm{\&alpha;}}_l\&CenterDot;s_l\sqrt{E_S}+n_l\left(k\right)---\left(20\right)$

Wherein, Filter output noise

is the Gaussian random variable of zero-mean, and corresponding variance is:

${\mathrm{\&sigma;}}_{n^{\&prime;}}^2=\underset{q}{\mathrm{\&Sigma;}}[E{\left|n(q;k)\right|}^2\&CenterDot;g(q;k)g^{*}(q;k)]$

$=\frac{N_0}{2}\&CenterDot;\left\{\underset{q}{\mathrm{\&Sigma;}}{\left|g(q;k)\right|}^2\right\}=\frac{N_0}{2}---\left(21\right)$

Know that by formula (15) each divides the noise component(s) n of comb formula subscriber signal gatherer output _l(k) independently of one another, therefore, the noise total amount that K branch comb formula subscriber signal gatherer introduced is K (N ₀/ 2), be the conventional filter noise intensity K doubly; But then, the signal summation of each minute comb formula subscriber signal gatherer collection is also corresponding to have increased K doubly, and corresponding signal energy is according to K ²Multiplication is long, and therefore extra reception processing gain has just been introduced in total seeing, we can see that systematic function has obtained obvious improvement really in the test of back.

Be similar to formula (7), we can provide subscriber signal gatherer k output noise component n _l(k) probability density function profiles:

$p[n_l\left(k\right)=n]=\frac{1}{\mathrm{\&pi;}{N}_0/2}\exp ({\left|n\right|}^2/(N_0/2))---\left(22\right)$

Wherein, n _l(k)=n representes the numerical value of the filtering output noise component of current branch comb formula subscriber signal gatherer, this moment d _l(k) conditional probability density also can correspondingly obtain:

$p\left(d_l\left(k\right)\right|s_l,{\mathrm{\&alpha;}}_l)=\frac{1}{\mathrm{\&pi;}{N}_0/2}\exp (-{|d_l\left(k\right)-{\mathrm{\&alpha;}}_l\&CenterDot;s_l|}^2/(N_0/2))---\left(23\right)$

Generally, the channel status α in the formula _lAll be to obtain through methods such as channel estimating at receiving terminal, and in order to accomplish processing such as signal equalization after this, might as well establish here that it is known to receiving terminal.Next, the corresponding soft information calculator that connects of branch comb formula subscriber signal gatherer k can be according to the d of current output _l(k) and channel status α _l, calculate the data symbol s of transmission _lGet a certain symbol of set Φ m ₁=0 ..., the soft value of declaring during M-1:

$P_k(s_l=s_{m_1})=\frac{p\left[d_l\left(k\right)\right|s_{m_1},{\mathrm{\&alpha;}}_l]}{\underset{s_{m^{\&prime;}}\&Element;\mathrm{\&Phi;}}{\mathrm{\&Sigma;}}p\left[d_l\left(k\right)\right|s_{m^{\&prime;}},{\mathrm{\&alpha;}}_l]}---\left(24\right)$

The system number that the communication system of representing M adopts, m ₁Be symbol number, s _{M '}Be the possible value of the institute of set Φ, following formula is illustrated in current channel condition α _lDivide comb formula subscriber signal gatherer k to receive signal d down, _l(k) judgement maybe value for symbol

Probability, existing with d _l(k) conditional probability density substitution following formula, can obtain branch road k output about data symbol s _lThe soft value of the declaring calculating formula of all feasible solutions:

$P_k(s_l=s_{m_1})=\frac{\exp ({|d_l\left(k\right)-{\mathrm{\&alpha;}}_l\&CenterDot;s_{m_1}|}^2/(N_0/2))}{\underset{s_{m^{\&prime;}}\&Element;\mathrm{\&Phi;}}{\mathrm{\&Sigma;}}\exp ({|d_l\left(k\right)-{\mathrm{\&alpha;}}_l\&CenterDot;s_{m^{\&prime;}}|}^2/(N_0/2))},m_1=0,...,M-1---\left(25\right)$

Then, branch road k will obtain about data symbol s _lSoft value of declaring of each feasible solution Deliver to decoding decision device module, and obtain the soft value of declaring together with other parallel branch and accomplish last symbol judgement, recover raw data symbols.

3.3. decoding judging module

The decoding judging module carries out comprehensively can adjudicating and recover original transmission symbol through soft value of declaring that each parallel branch of higher-dimension spatial noise signal demodulation process module is calculated.Here select the maximum a posteriori criterion as decision rule, decoder earlier with each branch road output about symbol s _lSoft value of declaring of each feasible solution add up respectively, that is:

$P(s_l=s_{m_1})=\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}[P_k(s_l=s_{m_1})/\underset{m_1=0}{\overset{M-1}{\mathrm{\&Sigma;}}}\underset{k=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{P}_k(s_l=s_{m_1})]---\left(26\right)$

Then selecting wherein, the pairing symbol value of maximum recovers raw data symbols as the optimal judgement result:

${\hat{s}}_l=\underset{s_{m_1}}{\arg }\left\{\underset{m_1=0,...,M-1}{\max }P(s_l=s_{m_1})\right\}---\left(27\right)$

Receiver of the present invention has been introduced independently random noise component of large-scale statistical, and makes each divide the noise component(s) n of comb formula subscriber signal gatherer output _l(k) independently of one another, this just makes systematic function improve becomes possibility.Therefore, the noise total amount introduced of K branch comb formula subscriber signal gatherer is K (N ₀/ 2), be the conventional filter noise intensity K doubly; But then, the signal summation that the subscriber signal gatherer on each parallel branch is collected is also corresponding to have increased K doubly, and corresponding signal energy is according to K ²Multiplication is long, and therefore extra reception processing gain has just been introduced in total seeing, can see that by embodiments of the invention systematic function has obtained obvious improvement really.

With specific embodiment realization of the present invention and effect are described below.

Instance 1: baseband transmission system

This example adopts bipolarity 2PSK modulation that data carry is transmitted on signal of communication.Character rate is the 1K baud, and the energy of single symbol is E _g, corresponding symbol is sent interval T ₀=1ms.The pulse shaping function that system adopts is that (length is T for Gauss's shaping function of unit energy ₀), that is:

$g\left(t\right)=\exp [-{(t-T_0/2)}^2/\left(2{\mathrm{\&sigma;}}_0^2\right)],t\&Element;[0,T_0]---\left(28\right)$

Wherein, parameter σ ₀=erfc (ρ)/2T ₀(ρ is the energy ratio).At this moment, system's originating terminal signal can be written as:

$s\left(t\right)=\sqrt{E_S}\&CenterDot;\underset{l}{\mathrm{\&Sigma;}}{s}_{l}g(t-l\&CenterDot;T)---\left(29\right)$

In the formula, s _lBe the data symbol that l time slot sends, get the 16QAM modulation in this example; Signal s (t) shared bandwidth B in process of transmitting is mainly determined by g (t).Here, energy is than ρ=10 ^-4, corresponding parameters value σ ₀=1.817 * 10 ^-4, its bandwidth that in transmission course, takies can calculate B=5.254KHz.

Receiving terminal adopts ultra wideband receiver structure of the present invention.In this example, the pass band width of its ultra broadband receiving filter is set to 3 times of common narrow band filter bandwidth, i.e. W ^Sup=15.762KHz has ([W ^Sup/ B]=3); Accordingly, the maximum multiple K of over-sampling sampling rate must not surpass 3, promptly

Choose by maximum in this example,

The corresponding sampling interval does

And standard sample is spaced apart T _S=0.1903ms.At this moment, the branch comb formula subscriber signal gatherer of receiving terminal has 3 the tunnel, is respectively g ₀, g ₁, g ₂, concrete form is following:

$\left\{\begin{array}{c}{g}_0={\left\{g\left(t\right)\right\}}_{t=n\&CenterDot;T_S}\\ g_1={\left\{g(t-T_S^{\mathrm{Sup}})\right\}}_{t=n\&CenterDot;T_S}\\ g_2={\left\{g(t-2T_S^{\mathrm{Sup}})\right\}}_{t=n\&CenterDot;T_S}\end{array}\right.,n=\mathrm{0,1},...,4---\left(30\right)$

Below the performance of this baseband transmission system under the flat fading wireless channel tested, and provide test result.Test condition is following: the channel maximum multipath time delay is 10.5ns, and maximum Doppler expands to 129.099Hz, and corresponding channel coherence time is 0.015s.Experiment is tested the performance of the ultra wideband receiver structure (2,3 times of standard bandwidths) that the present invention proposes; And compare with the receptivity of traditional narrow receiver; Like Fig. 6; " plus sige-dotted line " among the figure represented ultra wideband receiver of the present invention (2 times of standard bandwidths), and " triangle-chain-dotted line " represented ultra wideband receiver of the present invention (3 times of standard bandwidths), " circle-solid line " expression traditional narrow receiver.See that from test result the present invention can make receptivity obtain improving comparatively significantly under the same conditions.

Instance 2: orthogonal multiple carrier ofdm system

This example adopts orthogonal multiple carrier OFDM to accomplish the concurrent transmission of data, and the available carrier wave number is got Q=32 here.If complete OFDM mark space T ₀=1ms, subcarrier spacing are f _Δ(=1/T ₀)=1KHz.System adopt length be the rectangular pulse of 1ms as shaping function:

$g_q\left(t\right)=\frac{1}{\sqrt{T_0}}\exp \left[j2\mathrm{\&pi;}\left({\mathrm{qf}}_{\mathrm{\&Delta;}}\right)t\right],t\&Element;[0,T_0)q,q\&Element;[0,...,31]---\left(31\right)$

At this moment, system's originating terminal signal can be written as:

$s\left(t\right)=\&CenterDot;\underset{l}{\mathrm{\&Sigma;}}\underset{q=0}{\overset{Q-1}{\mathrm{\&Sigma;}}}{s}_{l,q}\sqrt{\frac{E_S}{T_0}}\exp \left[j2\mathrm{\&pi;}\frac{q}{T_0}(t-T_0)\right]---\left(32\right)$

Wherein, E _SBe the energy of single symbol, s _{L, q}Be the data symbol that sends on l the time slot q number of sub-carrier, its modulation system that adopts is 4PSK:

$s_{l,q}\&Element;\{(1\&PlusMinus;j)/\sqrt{2},(-1\&PlusMinus;j)/\sqrt{2}\}---\left(33\right)$

The bandwidth that this system takies in transmission course is B=Qf _Δ=32KHz.The sampling rate of common receiver is F _S=32KHz, corresponding sampling interval T _S=31.25 μ s; If receiving terminal then adopts the ultra wideband receiver structure, in this example, the pass band width of ultra broadband receiving filter is chosen according to 3 times of narrow band filter bandwidth equally, then W ^Sup=96KHz, corresponding high speed over-sampling speed does $F_S^{\mathrm{Sup}}=3F_S=96\mathrm{KHz},$ Sampling interval doesPromptly

Corresponding to each the subcarrier f that makes a start _q=qf _Δ(q=0 ..., 31), receiving terminal adopts with it corresponding 3 the tunnel to divide comb formula subscriber signal gatherer to accomplish the reception of signal respectively, is designated as g successively ₀(n; f _q), g ₁(n; f _q), g ₂(n; f _q), concrete form is following:

$\left\{\begin{array}{c}{g}_{q,0}\left(n\right)=\frac{1}{\sqrt{T}}\exp \left[j2\mathrm{\&pi;}\left({\mathrm{qf}}_{\mathrm{\&Delta;}}\right)\left(t\right)\right]{|}_{t=n{T}_S}\\ g_{q,1}\left(n\right)=\frac{1}{\sqrt{T}}\exp \left[j2\mathrm{\&pi;}\left({\mathrm{qf}}_{\mathrm{\&Delta;}}\right)(t-T_S^{\mathrm{Sup}})\right]{|}_{t=n{T}_S}\\ g_{q,2}\left(n\right)=\frac{1}{\sqrt{T}}\exp \left[j2\mathrm{\&pi;}\left({\mathrm{qf}}_{\mathrm{\&Delta;}}\right)(t-2\&CenterDot;T_S^{\mathrm{Sup}})\right]{|}_{t=n{T}_S}\end{array}\right.---\left(34\right)$

Test in the face of the performance of ofdm system (sub-carrier number Q=32) under the flat fading wireless channel down, and provide test result.Test condition is following: the OFDM symbol transmission interval is T=1ms, and subcarrier spacing is f _Δ(=1/T)=1KHz.It is T that system adopts length ₀The rectangular pulse of=1ms is as its shaping function.The channel maximum multipath time delay is 10.5ns, and maximum Doppler expands to 129.099Hz, and corresponding channel coherence time is 0.015s.Experiment is tested the performance of the ultra wideband receiver structure (3 times of standard bandwidths) that the present invention proposes; And compare with the receptivity of traditional narrow receiver; Like Fig. 7, " triangle-chain-dotted line " expression receiver of the present invention among the figure, " circle-solid line " expression traditional narrow receiver.See that from test result the present invention can make receptivity obtain improving comparatively significantly under the same conditions.

The present invention can make the system transmissions performance significantly improve on the basis of existing communication system, and has favorable applicability, is applicable to existing various types of communication system, has the potentiality and the value of extensive use.

Compare with communication system in the past, the present invention's unique distinction the most just is that it does not repel unwanted signal, but initiatively will more be with outer unwanted signal component to introduce into receiver, in order to participate in and the resuming work of completion data message.Concrete characteristics can be summarized as following two aspects again:

One of which; In ultra broadband receiving filter part; We mainly adopt the ultra high speed signal sampling to accomplish; The setting of its sampling rate

will be far above the shared bandwidth B of originating terminal signal reality, and

this point also is far different with the existing communication system;

Its two, the additional noise of introducing and unknown signaling component can be sent into demodulation and the special processing that higher-dimension spatial noise processing module is accomplished signal together with data-signal, and the soft-decision-value of dateout symbol, in order to accomplish final decoding judgement.In this course, with bringing extra processing gain, thereby make the performance of communication system obtain lifting by a larger margin.

Claims (4)

1. ultra-broadband receiver for digital communication system; It is characterized in that comprising the ultra broadband receiver module, higher-dimension spatial noise signal demodulation process module and the decoding judging module that connect successively; The ultra broadband receiver module comprises receiving filter and high speed over-sampling device; Said receiving filter bandwidth is for sending more than 2 times of signal bandwidth, and receiving filter exports high speed over-sampling device to, and the sampling rate of high speed over-sampling device is identical with the bandwidth of receiving filter; Higher-dimension spatial noise signal demodulation process module comprises a plurality of parallel branch comb formula subscriber signal gatherers; Each divides the output of comb formula subscriber signal gatherer to be connected with soft information calculator; The output of all soft information calculator is connected to the decoding judging module; Wherein dividing comb formula subscriber signal gatherer is the filter that is complementary with the transmitting terminal signal, and high speed over-sampling device exports said parallel branch comb formula subscriber signal gatherer to.

2. ultra-broadband receiver for digital communication system according to claim 1 is characterized in that branch comb formula subscriber signal gatherer is a comb filter.

3. the signal processing method of each described ultra-broadband receiver for digital communication system of claim 1-2; It is characterized in that receiving terminal in communication system; At first the receiving filter by the ultra broadband receiver module receives transmitting terminal signal and the outer noise cancellation signal of disturbing of transmitting terminal signal band in the lump; Again the signal that receives is carried out over-sampling; Resulting oversampled signals is delivered to higher-dimension spatial noise demodulation module and is done further processing, and after receiving oversampled signals stream, a plurality of parallel branch comb formula subscriber signal gatherer of higher-dimension spatial noise demodulation module is accomplished the multidiameter delay of signal and handled; And declare information and accomplish the decoding judgement to deciphering judging module by the output of soft information calculator is soft, recover original data message;

Over-sampling in the ultra broadband receiver module is:

It is that the transmission signal of B carries out broadband reception to bandwidth that the ultra broadband receiver module adopts receiving filter, and wherein B is for sending the bilateral spectral width of signal equivalence baseband signal, and the pass band width of receiving filter is W ^Sup, be the actual transmission signal bandwidth B K doubly, i.e. W ^Sup=KB, K>=2; Next, high speed over-sampling device carries out over-sampling to filter output signal, and sampling rate does The corresponding sampling interval It is standard nyquist sampling interval T _SThe 1/K of=1/B, promptly If the mark space of signal of communication is T ₀, then signal sampling point can be written as on l mark space:

Wherein, g () is for sending the pulse shaping function that symbol adopted, and p representes sample number, p=0 ..., K [T ₀/ T _S]-1 is promptly at mark space T ₀On adopt total sample be K [T ₀/ T _S], α _lBe the channel fading factor, s _l∈ Φ is a certain data symbol that l time slot sends, and Φ is data symbol s _lMight value all, E _SBeing the energy of single symbol, is W for pass band width ^SupFilter, the noise signal component n ' of its output (t) is uniformly-spaced putting t=p/W ^SupThe sampling point at place

Be the Gaussian random variable of separate same distribution, average is 0, and variance is N ₀/ 2, N ₀/ 2 is the bilateral power spectral density of noise;

Delivering to higher-dimension spatial noise signal demodulation module according to the oversampled signals sample that obtains after sampling rate

sampling handles; Higher-dimension spatial noise signal demodulation module correspondence is provided with K parallel branch comb formula subscriber signal gatherer, collects and screening operation in order to accomplish useful signal; Divide comb formula subscriber signal gatherer respectively oversampled signals to be carried out extracted at equal intervals according to different separately extraction of example side-play amounts; And carry out parallel processing to extracting signal; For the branch comb formula subscriber signal gatherer that is numbered k; Branch number of living in is k, and corresponding sample extracts side-play amount, and its extraction result is for

:

Wherein, n () expression receiving terminal noise and unknown signaling disturb, q=0 ..., [T ₀/ T _S]-1 is for extracting the numbering of sample value, the extraction sampling point that following formula obtains is actual be original signal r (t) with

As its initial offset time, according to standard nyquist sampling interval T _SThe result who samples is known r (q by sampling thheorem; K) can recover primary signal r (t) equally, the useful signal component that each road divides the extraction result of comb formula subscriber signal gatherer to be comprised is of equal value fully, and has undistortedly:

K wherein _iAnd k _jRepresent different branch numbers, promptly the useful signal energy collected of each minute comb formula subscriber signal gatherer all equates, and equals to send signal energy, divides comb formula subscriber signal gatherer k to use and extracts signal point range { r (q; K) } _qIn the pulse that is complementary of useful signal

Accomplish and extract the signal filtering reception, its output result is:

Wherein,

is for dividing the filtering output noise component of comb formula subscriber signal gatherer; Be the Gaussian random variable of zero-mean, this variable corresponding variance is:

g ^*() is the conjugate function of g (),

So, the output noise component n of branch comb formula subscriber signal gatherer k _l(k) probability density function profiles is:

Wherein, n _l(k)=n representes the numerical value of the filtering output noise component of current branch comb formula subscriber signal gatherer, this moment d _l(k) conditional probability density also can correspondingly obtain:

Next, the corresponding soft information calculator that connects of branch comb formula subscriber signal gatherer k can be according to the d of current output _l(k) and the channel fading factor-alpha _l, calculate the data symbol s of transmission _lGet a certain symbol of set Φ

m ₁=0 ..., the soft value of declaring during M-1:

The system number that the communication system of representing M adopts, m ₁Be symbol number, s _{M '}Be the possible value of the institute of set Φ, following formula is illustrated in current channel condition α _lDivide comb formula subscriber signal gatherer k to receive signal d down, _l(k) judgement maybe value for symbol

Probability, existing with d _l(k) conditional probability density substitution following formula, can obtain branch road k output about data symbol s _lThe soft value of the declaring calculating formula of all feasible solutions:

Then, branch road k will obtain about data symbol s _lSoft value of declaring of each feasible solution

Deliver to decoding decision device module, and obtain the soft value of declaring together with other parallel branch and accomplish last symbol judgement, recover raw data symbols.

4. the signal processing method of ultra-broadband receiver for digital communication system according to claim 3; It is characterized in that deciphering judging module and select the maximum a posteriori criterion as decision rule, earlier will by higher-dimension spatial noise signal demodulation process module obtain about data symbol s _lSoft value of declaring of each feasible solution add up respectively, that is:

Then selecting wherein, the pairing symbol value of maximum recovers raw data symbols as the optimal judgement result:

Images